Context: Building virtual databases of cell models with the so-called population of models (POM) approach is becoming a widely adopted in silico procedure, to account for the inherent variability of the cells within the cardiovascular system. Despite its popularity, POM implementations have remained as in-house tools, with no open-source tools available for the community.
The challenge with the POM approach is that it requires a complex and detailed set of steps that commonly fall upon each individual researcher to implement independently. This is not only inefficient but leaves significant room for mistakes and misunderstandings. These steps include model parameter expansion, biomarker calculation, and calibration. To address this, we introduce POMtool: a Python-based open-source tool to generate and handle the entire process of population generation.
Methodology: POMtool handles the running of the model based on the user defining their model's command-line interface, exposing the model parameters, selecting biomarkers, and selecting valid ranges for calibration. Once configured, POMtool can create a population of models, parameterize the model, or perform a sensitivity analysis. POMtool also supports user-controlled task partitioning for distributed cluster execution.
Demonstration: We demonstrate the use of our POMtool to create a hypertrophic cardiomyopathy (HCM) human induced pluripotent stem cell-derived cardiomyocyte population of models based on our earlier Forouzandehmehr2024 MATLAB model, to match the range described in the literature. We demonstrate how using the POMtool in producing a population of models representing the variety in the HCM mutation.
Conclusions: The POMtool application is reproducible with an easily shareable configuration file. This process is repeatable regardless of the underlying model, the model implementation language, or the computational environment.